Hidden Markov models for burst error characterization in indoorradio channels

摘要

Many digital communication channels exhibit statisticalndependencies among errors. The design of error control schemes for suchnchannels and their performance evaluation is simplified if appropriatengenerative models of the overall communication link are available. Thisnpaper presents a new class of generative models based on theninterconnection of hidden Markov submodels parameterized by thenBaum-Welch algorithm. The method has some resemblance to thenwell-studied problem of speech recognition of isolated words; however,nin our approach, instead of dealing with words, one deals with errornbursts, and the final goal is to generate bursts rather than tonrecognize words. The proposed model is particularly suitable fornsimulating error profiles with long bursts, as is often the case innindoor radio channels, where the error-free gaps inside a burst arenheavily nonrenewal. The merits of the method are corroborated bynapplying the technique to two particular examples of indoorncode-division multiple-access (CDMA) radio links